Fast prediction of possible flood risk before or during a concrete extreme weather event is obviously important. The (big) data needed in such predictions is to a large extent available. For example, a detailed grid terrain model with more than 200 billion measurements across Denmark is available from SDFE as part of the government’s basic data program, and detailed rain and sea level forecast and real time event data are available from the Danish Meteorological Institute (DMI). However, most flood risk assessment is currently only done off-line on models of extreme events (such as uniform rain or sea-level rise events, or predictions of 5, 10, 50 or 100 year events) or on data for historic events, and not on real-time forecasts or data (at least not using very large and detailed terrain models). One main reason for this is that current flood risk models cannot be run fast enough in order to be relevant for real-time data.
The goal of this case project is to develop algorithms and systems for online flood risk assessment based on up-to-date forecast or real-time data, that is, for fast prediction of flood risk for real events. More precisely, the goal is to combine DMI sea level forecasts with SCALGO technology for screening of risk from rising sea levels. The DMI sea level forecasts, which are updated every 6 hours and giving hourly predictions 5 days ahead, have shown to be particularly accurate. Potentially, the SCALGO technology is efficient enough to be able to on-line produce flood risk predictions from the DMI forecasts using the detailed Denmark terrain model.
The Danish Meteorological Institute (DMI), SCALGO, Aarhus University, The Danish Agency for Data Supply and Efficiency (SFDE) and Central Denmark Region